Spammers solving A.I. Problems?

New Scientist has published an article about how spammers could be the key to solving A.I. problems and eventually passing the Turing test. The one thing they do not mention is that CAPTCHAs are all about computers detecting other computers, while the Turing test is about humans detecting computers.

AGI

So it seems some other people have also noticed that the field of A.I. has lost its direction and have proposed starting a new name for the old A.I., A.G.I (Artificial General Intelligence).

There has even been a first conference and a wiki.

True A.I.

I was talking with a friend the other day about free will and we reached the point where we thought that free will is currently missing from alot of A.I. research. I would not consider any A.I. to be true A.I. without the machine having free will. This then brought up how I believe that there is true randomness if and only if there is free will. And that a machine must have a fuel source of true randomness in order to have free will. However, this is a topic of another blog post that perhaps will someday be written. Here I wanted to talk about...What the heck happened to A.I. research!?

I think at some point, every researcher who starts in A.I. realizes that there just really isn't much funding for foundational true A.I. research. All of the so-called A.I. research is currently more about data-mining and adaptive systems. I believe this is because these areas have a clearer short-term potential economic gain.

Who is working on the real foundational questions concerning constructing a true A.I.?